The role of Na⁺ channels. A, Current PSDs are shown for the sample cell in Figure 4 (at -50, -60, -70, and -80 mV) before and after the addition of TTX to the bath. There is no significant difference at -70 or -80 mV. The most consistent effect across all cells occurred at frequencies >100 Hz for holding potentials of -60 and -50 mV: the noise decreased with the blockage of Na⁺ channels. This was in agreement with the low amplitude, yet flat noise spectrum expected from fast Na⁺ channels, as demonstrated by the model (Fig. 8 B, red). The amplitude decrease in this example, at -50 mV, was larger than average. The current PSD from an idealized RC cell is shown as well, indicating the presence of instrumental noise at low amplitudes and at frequencies >100 Hz. B, The mean rms current noise was compared between control and TTX trials for both the sample cell (open circle) and the population (solid curves). The difference in the rms noise became increasingly significant with depolarization (p = 0.17, 0.28, 0.0019, and 7 × 10^-4; t test at -80, -70, -60, and -50 mV). C, In a t test comparison with and without TTX, the mean rms voltage noise was similarly affected by depolarization (p = 0.072, 0.068, and 0.012; t test at -80, -70, and -60 mV). Additional depolarization under current clamp evoked trains of action potentials, vitiating noise measurements. D, The slope n_I, of the best fit of A_I/1 + (f/f_c)^nI to the current PSD on double logarithmic coordinates, is shown averaged over the experimental data. Depolarization in the presence of TTX uncovered a voltage-dependent 1/f² noise component, most likely from the slow K⁺ current. t test comparisons of the slope in control versus TTX trials at each holding potential yielded (p = 0.19, 0.25, 7 × 10^-10, and 3 × 10^-6; t test at -80, -70, -60, and -50 mV).